• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 34 subscribers
  • Views 276 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

DP83867CS: DP83867CSRGZR Design question, PCB-trace link as well as auto-polarity correction in Mirror Mode

Isaac B
Prodigy 20 points

Part Number: DP83867CS
Other Parts Discussed in Thread: DP83869

Hello, I am designing an embedded PoE switch and have selected DP83867CSRGZR for two ports which require 1GBase-T. I have two of the PHYs on the board, and two questions.

 

  1. For one of the links, I want to use Mirror Mode for routing purposes. In the datasheet, it notes that in Mirror Mode, the polarity of the signals is reversed. In my layout, the receiver PHY has the PN signals swapped relative to the DP83867.

    In this case, Would the DP83867 support auto-polarity correction for the lanes? So in Mirror mode I will swap the pair orientation, but potentially keep the polarities the same?

    image.png
  2. I have seen a few forum posts related to this but I was wondering if there has been any updated testing or new guidance from TI. One of the ethernet links on my board necessitates having a PHY-to-PHY connection on the PCB traces (a few inch). I am wondering now about whether 1x 1:1 transformer is possible for this application, or whether you need two back-to-back, to simulate the cabled connection exactly. This is the current schematic: image.png
    This thread states that this method should be ok, but I just wanted to see if there was any updated stance on this https://e2e.ti.com/support/interface-group/interface/f/interface-forum/956195/dp83867ir-connect-2-phy-togeter-on-the-same-pcb

    One difference to the above is that my application will use a different PHY on the other end, instead of 2x DP83867 devices. It is a SOM with a built in Qualcomm AR8031-AL1A, which I cannot change at this time. Do you see any issues with using this connection scheme?

Thank you

  • 0
    TI__Mastermind 37565 points

    Hi Isaac,

    1. Yes the DP83867 will correct for polarity in mirror mode as well. It is ok to swap the P/N lines if necessary.

    2. Based on the previous E2E post, it may be possible to do this with one transformer between both PHYs. I agree with you that two transformers back-to-back would be a safer option since it more closely resembles a cabled connection. I'd recommend using two transformers, especially with a different PHY as the link partner.

    • Does the Qualcomm PHY support AC coupling on the MDI lines? It may be simpler to switch from DP83867 to DP83869 and use a set of AC coupling capacitors rather than transformers.

    Best,

    Shane

  • 0 in reply to Shane Hauser
    Prodigy 20 points

    Hi Shane,

    Thanks for your response. I am not sure about the QC PHY supporting AC coupling. That was originally the idea for the on-board link, but it seems like capacitive coupling tends to be tricky to implement in general and I figured the transformer option would be lower risk. 

  • +1 in reply to Isaac B
    TI__Mastermind 37565 points

    Hi Isaac,

    If you're not sure whether the QC PHY supports cap coupling, I agree its better to move forward with transformers.

    In this case, using DP83867 is ok.

    Best,

    Shane